Tooth microwear occurs when an animal processes food, producing microscopic pits and scratches on tooth surfaces, providing evidence of tooth movements and food properties. 3D microwear analysis is a growing field of study, where sub-micron scale tooth surface textures are used to compare populations with dietary differences. It has been primarily employed to study terrestrial vertebrates, however, the technique has rarely been applied to aquatic vertebrates, and never to aquatic mammals. Furthermore, the technique suffers from methodological variability. To address these points this thesis presents the results of five studies using 3D microwear analysis, three of which investigate different aspects of methodological variability, and two investigate the utility of 3D microwear analysis to differentiate diet in marine mammals, both across multiple species, and within a single species. An investigation of seven commonly used moulding compounds of varying viscosity demonstrated that mid-viscosity President Jet Regular Body produced the most accurate and precise moulds of tooth surface texture. An investigation was also carried out to test the effect of scale limiting 3D surfaces using 40 different combinations of operator (Nth order of polynomial) and filter to produce roughness surfaces. It was shown that high variability exists between resulting surfaces depending on the operator and the filter used. A combination of 6th order of polynomial, robust Gaussian filter and 0.025mm nesting index produced the greatest separation of known dietary groups while also being comparable to surfaces generated using many other combinations. An investigation into the effect of using four different microscopes to collect 3D tooth surface texture data showed high variability between resulting roughness parameter values and sensitivity to dietary differences depending on the microscope used. When testing the ability of 3D microwear analysis to separate ten marine mammal species into four known dietary groups, it was shown that this technique is highly sensitive to dietary differences, and provides information about the dietary evolution of extinct cetaceans. Finally, when using dentin tooth surfaces to test the ability of 3D microwear to detect differences between Orcinus orca dietary populations, it was found that their surface texture appears highly variable, and that little separation was possible between dietary groups.